BOLE: A script-knowledge driven multi-agent framework for reproducible llm-assisted genomic breeding

doi:10.1016/j.jia.2026.07.011

Advanced Online Publication | Current Issue | Archive | Adv Search

Zishuai Wang^1,2^,3*, Chongxiao Liang¹^*, Yanlin Zhang⁴, Rong Zhou⁵, Xiaoai Zhang², Wenkang Wei², Kui Li¹^#

1. Shenzhen Branch, Guangdong Laboratory of Lingnan Modern Agriculture/Key Laboratory of Livestock and Poultry Multi-omics of MARA, Agricultural Genomics Institute at Shenzhen, Chinese Academy of Agricultural Sciences, Shenzhen 518000, China

2. Agro-biological Gene Research Center/State Key Laboratory of Swine and Poultry Breeding Industry, Guangdong Academy of Agricultural Sciences, Guangzhou 510640, China

3. Agsino GenSources Co.,Ltd., Shenzhen 518000, China

4. Data Science and Analytics Thrust, Hong Kong University of Science and Technology (Guangzhou), Guangzhou 510145, China

5. The State Key Laboratory of Animal Biotech Breeding, Institute of Animal Sciences, Chinese Academy of Agricultural Sciences, Beijing 100193, China

Highlights

Presents an intelligent orchestration framework for scalable genomic breeding analysis.

Balances LLM-driven autonomy with knowledge-based constraints to ensure robustness.

Cross-task integration of GWAS, heritability, germplasm evaluation, and genomic selection.

Establishes a reproducible, intent-driven alternative to static bioinformatics pipelines.

Abstract
References

Download: PDF in ScienceDirect
Export: BibTeX | EndNote (RIS)

摘要

基因组育种已迈入高度数据密集型时代，然而，将海量异质生物信息学工具整合为连贯、可复用的分析工作流，仍是制约该领域发展的核心瓶颈。软件依赖的复杂性、数据格式的多样性以及分析流程的专业门槛，使得基因组分析的可及性与可重复性大打折扣，尤其对非专业用户而言更是如此。针对上述挑战，我们开发了 BOLE（伯乐）——一个以知识驱动的多智能体AI框架，通过约束大语言模型推理来实现基因组育种分析的自动化与标准化。伯乐将分析模块形式化为脚本单元，每个单元均明确定义输入输出与软件依赖规范，从而实现对生物信息学工作流的系统化组织。框架内置基于模式定义的验证机制，对大语言模型生成的分析方案与可执行代码进行约束，确保工作流生成的可靠性与一致性。伯乐覆盖了基因组育种的核心分析任务，包括全基因组关联分析、遗传力估计、种质资源评估及基因组选择等。与此同时，框架完整保留分析产物层面的数据溯源信息，从根本上保障了分析结果的可重复性。伯乐以网站服务形式部署，支持从原始数据到可操作结果的端到端基因组育种分析。通过提升工作流自动化水平、降低使用门槛并确保分析可重复，伯乐为人工智能与基因组育种的深度融合贡献了一种可推广的范式。框架已开源发布，访问地址为 https://bole.zishuailab.com/。

Abstract

Genomic breeding has become increasingly data-intensive, yet the practical integration of heterogeneous bioinformatics tools into coherent analytical workflows remains a major bottleneck. The complexity of software dependencies, data formats, and analytical procedures often limits the accessibility and reproducibility of genomic analyses, particularly for non-expert users. To address these challenges, we developed BOLE, a structured script-knowledge driven multi-agent AI framework with constrained large language model reasoning for genomic breeding. BOLE formalizes analytic modules as script units with explicit input-output contracts and dependency specifications, enabling the systematic organization of bioinformatics workflows. The framework incorporates schema-based validation to constrain large language model generated analysis plans and executable code, ensuring reliability and consistency in workflow generation. BOLE supports major genomic breeding analyses, including genome-wide association studies, heritability estimation, germplasm evaluation, and genomic selection, while preserving artifact-level provenance to guarantee reproducibility. Implemented as a web-based service, BOLE enables end-to-end genomic breeding analysis from raw data to actionable results. By improving workflow automation, accessibility, and reproducibility, this framework offers a scalable paradigm for integrating artificial intelligence into genomic breeding research and practice. BOLE is freely available at https://bole.zishuailab.com/.

Keywords: genomic breeding multi-agent large language model GWAS genomic selection

Online: 10 July 2026

Fund:

This work was supported the National Key Research & Developmental Program of China (2021YFF1000600), the Special Funding for the Construction of the High-Level Academy of Agricultural Sciences, China (NYQS202632), the Fund of State Key Laboratory of Swine and Poultry Breeding Industry, China (ZQQZ-G20, ZQQZ-G22), the Shenzhen Outstanding Talents Training Fund, China (202102) and Guangdong Provincial Project, China (2024QN11N085).

About author: Zishuai Wang, E-mail: wangzishuai@gdaas.cn; Chongxiao Liang, E-mail: 821012450063@caas.cn; #Correspondence Kui Li, E-mail: likui@caas.cn * These authors contributed equally to this work.

	Service
	E-mail this article
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors

Cite this article:

Zishuai Wang, Chongxiao Liang, Yanlin Zhang, Rong Zhou, Xiaoai Zhang, Wenkang Wei, Kui Li. 2026. BOLE: A script-knowledge driven multi-agent framework for reproducible llm-assisted genomic breeding. Journal of Integrative Agriculture, Doi:10.1016/j.jia.2026.07.011

Afgan E, Baker D, Batut B, Van Den Beek M, Bouvier D, Čech M, Chilton J, Clements D, Coraor N, Grüning B A. 2018. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update. Nucleic Acids Research, 46, W537-W544.

Alexander D H, Novembre J, Lange K. 2009. Fast model-based estimation of ancestry in unrelated individuals. Genome Research, 19, 1655-1664.

Atkinson M, Gesing S, Montagnat J, Taylor I. 2017. Scientific workflows: Past, present and future. Future Generation Computer Systems,75, 216-227.

Bai J, Bai S, Chu Y, Cui Z, Dang K, Deng X, Fan Y, Ge W, Han Y, Huang F. 2023. Qwen technical report. ArXiv, 2309,16609.

Bao J, Xie Q. 2022. Artificial intelligence in animal farming: A systematic literature review. Journal of Cleaner Production, 331, 129956.

Bi X, Chen D, Chen G, Chen S, Dai D, Deng C, Ding H, Dong K, Du Q, Fu Z. 2024. Deepseek llm: Scaling open-source language models with longtermism. ArXiv,2401,02954.

Di Tommaso P, Chatzou M, Floden E W, Barja P P, Palumbo E, Notredame C. 2017. Nextflow enables reproducible computational workflows. Nature Biotechnology, 35, 316-319.

Feng X-K, Xie C-D, Li Y-Y, Wang Z-S, Bai L-J. 2023. SCSMRD: a database for single-cell skeletal muscle regeneration. Journal of Integrative Agriculture, 22, 864-871.

Gianola D, De Los Campos G. 2008. Inferring genetic values for quantitative traits non-parametrically. Genetics Research, 90, 525-540.

Habier D, Fernando R L, Kizilkaya K, Garrick D J. 2011. Extension of the Bayesian alphabet for genomic selection. BMC Bioinformatics, 12, 186.

Hayes B, Goddard M. 2010. Genome-wide association and genomic selection in animal breeding. Genome, 53, 876-883.

He X, Li Z, Guo S, Zheng X, Liu C, Liu Z, Li Y, Yuan Z, Li L. 2024. Epistasis-aware genome-wide association studies provide insights into the efficient breeding of high-yield and high-quality rice. Journal of Integrative Agriculture, 23, 2541-2556.

Huang C, Zheng M, Huang Y, Cai L, Zou X, Yao T, Xie X, Yang B, Xiao S, Ma J, Huang L. 2026. Unraveling genetic underpinnings of purine content in pork. Journal of Integrative Agriculture, 25, 1099-1113.

Ji Z, Lee N, Frieske R, Yu T, Su D, Xu Y, Ishii E, Bang Y J, Madotto A, Fung P. 2023. Survey of hallucination in natural language generation. ACM Computing Surveys, 55, 1-38.

Leipzig J. 2017. A review of bioinformatic pipeline frameworks. Briefings in Bioinformatics, 18, 530-536.

Leitão P. 2009. Agent-based distributed manufacturing control: A state-of-the-art survey. Engineering Applications of Artificial Intelligence, 22, 979-991.

Lewis P, Perez E, Piktus A, Petroni F, Karpukhin V, Goyal N, Küttler H, Lewis M, Yih W-T, Rocktäschel T. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems, 33, 9459-9474.

Li K, An B, Liang M, Chang T, Deng T, Du L, Cao S, Du Y, Li H, Xu L. 2024. Prescreening of large-effect markers with multiple strategies improves the accuracy of genomic prediction. Journal of Integrative Agriculture, 23, 1634-1643.

Liang J, Huang W, Xia F, Xu P, Hausman K, Ichter B, Florence P, Zeng A. 2022. Code as policies: Language model programs for embodied control. ArXiv, 2209, 07753.

Liu G, Ren W, Jin K, Zheng D, Zuo Q, Zhang Y, Chen G, Li B, Niu Y. 2025. PGC-mediated conservation strategies for germplasm resources of Rugao Yellow chicken and Shouguang chicken in China. Journal of Integrative Agriculture, 24, 2327-2341.

Lubanovic B. 2023. FastAPI. O'Reilly Media, California, United States. pp. 27-47.

Mehandru N, Hall A K, Melnichenko O, Dubinina Y, Tsirulnikov D, Bamman D, Alaa A, Saponas S, Malladi V S. 2025. BioAgents: Bridging the gap in bioinformatics analysis with multi-agent systems. Scientific Reports, 15, 39036.

Meuwissen T H, Hayes B J, Goddard M. 2001. Prediction of total genetic value using genome-wide dense marker maps. Genetics, 157, 1819-1829.

Patterson H D, Thompson R. 1971. Recovery of inter-block information when block sizes are unequal. Biometrika, 58, 545-554.

Pérez P, De Los Campos G. 2014. Genome-wide regression and prediction with the BGLR statistical package. Genetics, 198, 483-495.

Piccolo S R, Ence Z E, Anderson E C, Chang J T, Bild A H. 2021. Simplifying the development of portable, scalable, and reproducible workflows. Elife, 10, e71069.

Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira M A, Bender D, Maller J, Sklar P, De Bakker P I, Daly M J. 2007. PLINK: a tool set for whole-genome association and population-based linkage analyses. The American Journal of Human Genetics, 81, 559-575.

Reiter T, Brooks P T, Irber L, Joslin S E, Reid C M, Scott C, Brown C T, Pierce-Ward N T. 2021. Streamlining data-intensive biology with workflow systems. GigaScience, 10, giaa140.

Rexroad C, Vallet J, Matukumalli L K, Reecy J, Bickhart D, Blackburn H, Boggess M, Cheng H, Clutter A, Cockett N. 2019. Genome to phenome: improving animal health, production, and well-being–a new USDA blueprint for animal genome research 2018–2027. Frontiers in Genetics, 10, 327.

Sandve G K, Nekrutenko A, Taylor J, Hovig E. 2013. Ten simple rules for reproducible computational research. PLoS Computational Biology, 9, e1003285.

Shen Y, Song K, Tan X, Li D, Lu W, Zhuang Y. 2023. Hugginggpt: Solving ai tasks with chatgpt and its friends in hugging face. Advances in Neural Information Processing Systems, 36, 38154-38180.

Su H, Long W, Zhang Y. 2025. BioMaster: Multi-agent System for Automated Bioinformatics Analysis Workflow. BioRxiv, 2023,634608.

Vanraden P M. 2008. Efficient methods to compute genomic predictions. Journal of Dairy Science, 91, 4414-4423.

Wang J, Zong W, Shi L, Li M, Li J, Ren D, Zhao F, Wang L, Wang L. 2026. Using mixed kernel support vector machine to improve the predictive accuracy of genome selection1. Journal of Integrative Agriculture , 25, 775-787.

Wang X, Li X, Chen D, Gao J, Hao S, Zhang H, Zhao Z, Shen M, Chen H, Qi F, Zhang K, Zhou H, Xi Y, Zhou J, Yu Y, Xu Q. 2026. Core germplasm construction of tea plant populations based on genome-wide SNP and catechins in Shaanxi Province, China. Journal of Integrative Agriculture, 25, 1927-1938.

Wang Z, Li W, Tang Z. 2024. Enhancing the genomic prediction accuracy of swine agricultural economic traits using an expanded one-hot encoding in CNN models1. Journal of Integrative Agriculture, 24, 3574-3582.

Wooldridge M. 2009. An Introduction to Multiagent Systems. John wiley & sons, New York City, United States. pp. 1-27.

Xi Z, Chen W, Guo X, He W, Ding Y, Hong B, Zhang M, Wang J, Jin S, Zhou E. 2025. The rise and potential of large language model based agents: A survey. Science China Information Sciences, 68, 121101.

Xue B, Xin L, Wei-Jia W, Zi-Wei L, Ning W, Fan X, Hai-He G, Huai-Shun G, Hui L, Shou-Zhi W. 2022. Integration of genome-wide association study and selection signatures reveals genetic determinants for skeletal muscle production traits in an F2 chicken population. Journal of Integrative Agriculture, 21, 2065-2075.

Yang J, Benyamin B, Mcevoy B P, Gordon S, Henders A K, Nyholt D R, Madden P A, Heath A C, Martin N G, Montgomery G W. 2010. Common SNPs explain a large proportion of the heritability for human height. Nature Genetics, 42, 565-569.

Yang J, Lee S H, Goddard M E, Visscher P M. 2011. GCTA: a tool for genome-wide complex trait analysis. The American Journal of Human Genetics, 88, 76-82.

Yang R, Guo X, Zhu D, Tan C, Bian C, Ren J, Huang Z, Zhao Y, Cai G, Liu D, Wu Z, Wang Y, Li N, Hu X. 2021. Accelerated deciphering of the genetic architecture of agricultural economic traits in pigs using a low-coverage whole-genome sequencing strategy. GigaScience, 10, giab048.

Zhang S, Hu X, Wang F, Miao H, Ye C, Yang W, Zhong W, Chen J. 2025. Identification of QTLs for plant height and branching-related traits in cultivated peanut. Journal of Integrative Agriculture, 24, 2511-2524.

No related articles found!

No Suggested Reading articles found!

Viewed

Full text

Abstract

Cited

Shared

Discussed

Cite this article:

About JIA

Editorial board

For authors